IRAQI JOURNAL OF STATISTICAL SCIENCES

Abstract

Three suggested procedures were adopted to determine the value of biasing parameter (k) in ridge regression: 1-fragmenting the ridge trace to groups each group contain semi-homogeneous absolute values of the estimated parameters, 2-rotating over the axis of ridge trace, 3-the ideal value of k should be determined through many analytical stages.
A procedure was suggested to select the best variables to admit in the ridge regression equation. The augmented matrix and vector have had a significant importance in facilitating all analyzing processes.